LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 1 Introduction to the Detector Characterization Sessions Keith Riles (University of Michigan) Daniel Sigg (LIGO–LHO) LIGO Scientific Collaboration Meeting LIGO Livingston Observatory March 18, 2003

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 2 Recent Progress Tremendous progress in last year on all fronts: –DMT infrastructure much enhanced (thanks to John, Szabi, Daniel, Masahiro) –Many DMT monitors reaching maturity (thanks to DaveC, Ed, Masahiro, Szabi, Sergey, Rauha, PatrickS, John) –New DMT monitors in the pipeline (thanks to TimO, AdrianO, BrianS, Natalia)

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 3 Recent Progress Tremendous progress in last year on all fronts: (cont.) –Data run operations much improved, after E7 / S1 troubles: Clearer (and more useful) scimon instructions Figures of merit More automation More projectors (thanks, Fred & Mark!) (thanks to many authors, testers, & scimons; M1, M2, E9 runs well worth the effort!) –Particularly impressive improvements in procedures for Calibrations (thanks to Gaby, Mike, Brian, Rana, PatrickS) Signal Injections (thanks to PeterSh, Szabi, Isabel, Sukanta)

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 4 Recent Progress Tremendous progress in last year on all fronts: (cont.) –Online diagnostics for DAQ / CDS state / LDAS (thanks to DaveB, Chethan, PeterSh, Greg, Igor) –Coupling of online calibration with astrophysical analysis (thanks to PatrickB, Jolien, Duncan) –Online feedback from astrophysical searches (thanks to Duncan, AlanWe, Laura) –Pool of scimon experts expanded dramatically in S2

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 5 Worries… Despite this progress, there are long-term worries: S2 Investigators, upper limits analysts, and commissioners need to do better in closing the loop:S2 Investigators S2 Investigation Teams Upper Limits Analysts Commissioners DMT Authors

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 6 Worries… Closing the loop: Need more systematic studies of artifacts: Compilation  Cataloguing  Identification Most important artifacts: Frequent Glitches & Strong Lines Biggest problem: Manpower (no shortage of tools to use!) Unambiguous glitch identification would help justify a priori vetoing of epochs in burst and inspiral searches Glitch and line identification would aid commissioners in reducing noise and aid DMT authors in devising / refining monitors

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 7 Worries… Closing the loop: Commissioners should look at data-run figures of merit between runs (like the inspiral range!) DMT authors should talk to commissioners to learn how to make monitors more useful during / between runs, and commissioners should talk to authors! Upper limits analysts should tell commissioners what instrumental troubles limit them most, including noise floor (e.g., stochastic figures of merit for different Ώ(f) shapes)

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 8 Worries… Related worry: More scientists needed at the site: »Still more than an order of magnitude to go to reach design, and progress is getting harder »Postdocs & graduate students in residence could help enormously both in creating software tools for commissioning and in improvement of data characterization  DMT potential has only begun to be tapped! »Unauthorized advertisement: The DMT is a user-friendly environment for algorithm development on real data. An observatory is a good place to meet the experts who know what all those auxiliary channels mean…

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 9 Worries… Another worry: Must (eventually) enter a continuous-data-taking mode »Detchar and Upper Limit groups must respond accordingly:  Ongoing characterization and analysis, in time to allow intervention, if needed  More automation needed in many places  But also more scientists looking at results »Scimon shifts more burdensome: 24/7 scimon staffing impractical(?)  Start relying on single-shift / day monitoring? Remote shifts?  Demands better alarms, better instructions, more knowledge than we have now  Demands more stable interferometers (H1 in S2 may be a good model to study)

LIGO-G Z Introd. to Detect. Charact. Sessions K. Riles - University of Michigan 10 This meeting Lots of activity: 38 presentations (detchar, detchar/asis) Looking forward to interesting reports!